*************************************************************************************
*IMPORTANT NOTICE: THE DATA SET CAN NOT BE USED FOR OTHER REASONS THAN TO REPLICATE
*THE RESULTS OF THE PAPER!!!!!!!!!!!!!!!!!!!!!
*************************************************************************************


Estimation done in stata 12.1; Operating system: Windows server 2012
(except for the OLS1 results in Table 2, the wage and SUR results in Table 7 and the results in Table 9. These were obtained using Stata 11.2; Operating system: Windows 7)

******************************************************************************
*PROGRAMS USED FOR ESTIMATION
******************************************************************************

The data file basefile.dta contains the data used for the estimation. This file is generated from the file masterfile.dta using the do-file prepare_data.do. The description of the variables is included in the labels and at the end of this file. 

The do-file polynomialv5_.ado contains a program to construct a polynomial of specified order from a list of variables.


Table 2:

	Do-files: 
		
		baseresults.do 		for columns OLS2 and ACF for the first 9 columns
		baseresults_OLS.do 	for columns OLS1
		baseresults_lag.do 	for last three columns (ACF, Lag Labor)

	Output-files:
		baseresults_total.log	for columns OLS2 and ACF (Total)
		baseresults_manuf.log	for columns OLS2 and ACF (Manufacturing)
		baseresults_nonmanuf.log for columns OLS 2 and ACF (Nonmanufacturing)
		baseresults_OLS_fullsample.log for column OLS 1 (Total)
		baseresults_OLS_manuf.log for column OLS 1 (Manufacturing)
		baseresults_OLS_nonmanuf.log for column OLS 1 (Nonmanufacturing)
		baseresults_lag.log 	for for last three columns (AFF, Lag labor)

Table 3:
	Do-files:
		bluewhite_imperf.do	generates results reported in Table 3
		
	Output files:
		bluewhite_imperf.log	results reported in Table 3

Tables 4 and 5:

	Do-files:
		
		baseresults_persector.do generates data file acf_train_bootstrapREstat.dta with the results per 						NACE 2 digit sector  

	Output files:
		baseresults_persector.log
		acf_train_bootstrapREStat.dta: stata file with results per NACE 2 digit sector



Table 6:
	Do-files:
		baseresults_cont.do	Performs both the estimation for the sectors pooled together and generates 					data file acf_train_bootstrapREstatcont.dta with the results per sector

	Output Files:
		baseresults_cont.log	
		acf_train_bootstrapREstatcont.dta: results per NACE 2 digit sector 	


Table 7:
	Do-files:
		
		furtherrobust_stock.do 	for "Training Stock" (first, the do-file construct_stock.do needs to be 						run to generate the data file trainingstock_tomerge.dta)
		furtherrobust_flex.do 	for "Fully Flexible Materials"
		furtherrobust_wage.do 	for "Wage as Control"
		furtherrobust_sur.do	for "SUR Model"
		furtherrobust_trainlag for "Lagged Training"


	Output files:
		furtherrobust_stock.log for "Training Stock"
		furtherrobust_flex.log 	for "Fully Flexible Materials"
		furtherrobust_wage.log	for "Wage as Control"
		furtherrobust_sur.log	for "SUR model"
		furtherrobust_trainlag.log for "Lagged Training"

Table 8:
	Do-files:
		
		hetero_school.do	for Schooling results
		hetero_bluewhite.do	for Type of Contract results
		hetero_blueschoool.do	for Type of Contract and Schooling results
		hetero_fem.do		for Female/Male Employees
		


	Output files
		hetero_school.log	for Schooling results
		hetero_bluewhite.log 	for Type of Contract results
		hetero_blueschool.log	for Type of Contract and Schooling results				
		hetero_fem.log		for Female/Male employees

Table 9:
	Do-file:
		turnover.do		generates resutls reported in table 9


	Log file:
		turnover.log


Figure 1:
	Do-file:
		figure1.do		

	Data-file:
		acf_train_bootstrapREStat.dta contains the sector level estimates 



The following do-files contain the programs to perform the estimation (these are called in the do-files described above):

	acf_train_bootstrapped.do	for base specification (tables 2, 4-5)
	acf_train_bootstrapped_lag	for specification with lags of labor as instruments (table 2)
	acf_train_bootstrapped_blue.do	for results imperfect substitution between blue and white collar workers
	acf_train_bootstrapped_stock	for specification with the stock of labor (Table 7)
	acf_train_bootstrapped_school	for specification with schooling (Table 8)
	acf_train_bootstrapped_bluewhite.do for specification with Type of Contract (Table 8)
	acf_train_bootstrapped_blueschool.do for specification with Type of Contract and Schooling (Table 8)
	acf_train_bootstrapped_female.do for specification with Male/Female employees (Table 8)


*********************************************************************
*PROGRAMS FOR RAUCH CLASSIFICATION AND CONSTRUCTION STOCK VARIABLE
*********************************************************************
	
The folder classification_inputs contains the programs to classify the sectors according to whether the inputs are more likely to be variable or not using the Rauch (1999) classification

	Data files:
		tradeBE.dta: 		trade data for Belgium from UN Comtrade
		sitc2isic.dta:		conversion table for SITC Rev. 2 to ISIC Rev. 2 (from Muendler, 2009)
		rauchclas.dta: 		Rauch (1999) classification at the 4 digit SITC Rev. 2 level
		use_Belgium2000.dta:	Use Table for Belgium (2000) from Eurostat

	Do-file:
		goodsclass.do:		results in an estimate the share of inputs that is differentiated, 							reference priced or homogeneous for each different sector (reported 							in Figure B.1. in the appendix)

The folder stock contains the programs to compute the training stock used in the paper

	Do-file: 
		construct_stock.do:	do-file to compute training stock under different assumptions for the 							discount rate. The execution of the program results in a datafile 							trainingstock_tomerge.dta that is needed for the results on the stock 							variable (cf. Table 7 "Training Stock")



*************************************************************************
*VARIABLE DESCRIPTION
*************************************************************************


tangibleassetstheur		Tangible Assets, thousand euros
rawmaterialsconsumables		Raw materials, consumables, thousand euros
infte				Entrants - full-time equivalents : nber of empl. entered in the staff register
inmprim   			Entrants - full-time equivalents : primary education (males)
inmsec          		Entrants - full-time equivalents : secundary education (males)
inmhigh         		Entrants - full-time equivalents : higher education (males)
inmuniv         		Entrants - full-time equivalents : university education (males)
infprim         		Entrants - full-time equivalents : primary education (females)
infsec         			Entrants - full-time equivalents : secundary education (females)
infhigh         		Entrants - full-time equivalents : higher education (females)
infuniv         		Entrants - full-time equivalents : university education (females)
outfte          		Departures - full-time equivalents : nber of empl. entered in the staff register
outmprim        		Departures - full-time equivalents : primary education (males)
outmsec         		Departures - full-time equivalents : secundary education (males)
outmhigh        		Departures - full-time equivalents : higher education (males)
outmuniv        		Departures - full-time equivalents : university education (males)
outfprim        		Departures - full-time equivalents : primary education (females)
outfsec         		Departures - full-time equivalents : secundary education (females)
outfhigh        		Departures - full-time equivalents : higher education (females)
outfuniv        		Departures - full-time equivalents : university education (females)
outret          		Departures - full-time equivalents : retirement
outearlyret     		Departures - full-time equivalents : early retirement
outdism         		Departures - full-time equivalents : dismissal
outoth          		Departures - full-time equivalents : other
outself         		Departures - full-time equivalents : self-employed, at least half-time
trainl         			Nber of empl. involved in training initiatives
trainh          		Nber of training hours
trainc          		Training costs to the company thousand euros
addedvalue      		Value Added, thousand euros
idnumber        		ID number
nace4           		NACE 4 digit code
nace3           		NACE 3 digit code
nace2           		NACE 2 digit code
ppi_PRIN        		Domestic output price index - in national currency, from eurostat
mark            		group(idnumber)
GO_P            		Gross output deflator from EU klems
capitaldefl     		Capital Deflator
avemplfull      		Average nbr. full-time empl.
staffcostfull   		Staff costs full-time empl. thousand euros
avemplpart      		Aver. nber of part-time empl.
staffcostpart   		Staff costs of part-time empl. thousand euros
avemplfte       		Aver. nber of full-time equivalents empl.
hoursfte        		Nber of hours actually worked by full-time equivalents empl.
staffcostfte    		Staff costs of full-time equivalents empl. thousand euros
endemplfte      		Nber of full-time equivalents empl. entered in the staff register
endemplmfte     		Nber of full-time equivalents empl. on the balancing date (males)
endemplffte     		Nber of full-time equivalents empl. on the balancing date (females)
endemplmanfte   		Nber of full-time equivalents empl. on the balancing date (Management staff )
endemplbedien		        Nber of full-time equivalents empl. on the balancing date (Employees)
endemplarbfte                   Nber of full-time equivalents empl. on the balancing date (Workers)
endemplothfte                   Nber of full-time equivalents empl. on the balancing date (Other)
incorpyear                      Year of incorporation

